"Automatically Connect" and Static vs Dynamic IP addresses

[darcyr]

Hello!

I have been encountering this strange issue where we would set a static IP on an ethernet port, and when we restart the device, it suddenly swaps over to dynamic / DHCP.

What could cause this?

I wanted to check in on the "Automatically Connect" checkbox - does this mean that it will *try* to connect using the given settings, or that it will try to connect with whatever setting is set, and if it fails, try with DHCP?

I have been getting very frustrated with some networking teams as we try to install devices, running into all kinds of IP address issues.

[jlehtone]

we set a static IP on an ethernet port, and when we restart the device, it suddenly swaps over to dynamic / DHCP.

How do you set the IP?


PS. My preferred method to set static IP is by DHCP server handing out predetermined address for the device.

[darcyr]

we set a static IP on an ethernet port, and when we restart the device, it suddenly swaps over to dynamic / DHCP.

How do you set the IP?


PS. My preferred method to set static IP is by DHCP server handing out predetermined address for the device.

We are setting it using `nmtui`.

For your preferred method to set static IP, does this mean that the eth device is set to DHCP but theoretically ends up receiving a static one consistently?

[jlehtone]

We are setting it using `nmtui`.

That should make it stick.

You can list config with 'nmcli'.

Code: Select all

nmcli      # shows summary
nmcli d s  # shows device status
nmcli c s  # shows connections

When you know the name of UUID of connection (shown with nmcli c s) you can look details of one connection:

Code: Select all

nmcli c s conn_name  # all config
nmcli -f ipv4 c s conn_name  # just the 'ipv4' nm-settings

Note that when address is set manually (ipv4.method manual ipv4.addresses 1.2.3.4/5) the system will first shout (broadcast) out "Does anyone have 1.2.3.4 in use?" so that it won't cause address collision.

For your preferred method to set static IP, does this mean that the eth device is set to DHCP but theoretically ends up receiving a static one consistently?

Yes. The clients use DHCP to get config from server.
The server hands out specific address based on identifier (MAC address) of the client. (And only MACs known to the server do get address.)

The default "home DHCP" is set to give next available address from a pool to anyone.
While even that might tend to hand same address, there is no guarantee. Receiving "random" address = "dynamic".
One can also have a "dynamic pool", but only for known clients.
How much DHCP server can be configured does depend on the server. Even the cheap ones have some options.

[darcyr]

These are the values being pulled. Is there anything suspicious here? The "enp4s0" profile just fell off from static to DHCP...

Code: Select all

[root@host ~]# nmcli -f ipv4 c s enp4s0
ipv4.method:                            manual
ipv4.dns:                               --
ipv4.dns-search:                        --
ipv4.dns-options:                       ""
ipv4.dns-priority:                      0
ipv4.addresses:                         10.120.78.50/23
ipv4.gateway:                           --
ipv4.routes:                            --
ipv4.route-metric:                      -1
ipv4.route-table:                       0 (unspec)
ipv4.routing-rules:                     --
ipv4.ignore-auto-routes:                no
ipv4.ignore-auto-dns:                   no
ipv4.dhcp-client-id:                    --
ipv4.dhcp-timeout:                      0 (default)
ipv4.dhcp-send-hostname:                yes
ipv4.dhcp-hostname:                     --
ipv4.dhcp-fqdn:                         --
ipv4.never-default:                     no
ipv4.may-fail:                          yes
ipv4.dad-timeout:                       -1 (default)

[root@host ~]# nmcli -f ipv4 c s "System eno1"
ipv4.method:                            manual
ipv4.dns:                               9.9.9.9,8.8.8.8
ipv4.dns-search:                        --
ipv4.dns-options:                       ""
ipv4.dns-priority:                      0
ipv4.addresses:                         192.168.0.245/24
ipv4.gateway:                           192.168.0.1
ipv4.routes:                            --
ipv4.route-metric:                      -1
ipv4.route-table:                       0 (unspec)
ipv4.routing-rules:                     --
ipv4.ignore-auto-routes:                no
ipv4.ignore-auto-dns:                   no
ipv4.dhcp-client-id:                    --
ipv4.dhcp-timeout:                      0 (default)
ipv4.dhcp-send-hostname:                yes
ipv4.dhcp-hostname:                     --
ipv4.dhcp-fqdn:                         --
ipv4.never-default:                     no
ipv4.may-fail:                          yes
ipv4.dad-timeout:                       -1 (default)

[jlehtone]

Looks "normal".

[darcyr]

Looks "normal".

So do you think it's possible the network itself is somehow forcing the port to use DHCP instead of the static IP that was set?

It only seems to do this *if* the DHCP connection will work and connect to the internet.

[jlehtone]

So do you think it's possible the network itself is somehow forcing the port to use DHCP instead of the static IP that was set?

No.